Mortality for individuals with familial hypercholesterolemia (FH) in the period 1992 to 2010

(1)

Mortality for individuals with familial

hypercholesterolemia (FH) in the period 1992 to 2010

Mirza Sarancic

Master of Science Thesis in Clinical Nutrition Department of Nutrition, Faculty of Medicine

UNIVERSITY OF OSLO May 2012

(2)

Mortality for individuals with familial

hypercholesterolemia (FH) in the period 1992 to 2010

A registry-based study

Master of Science Thesis in Clinical Nutrition Mirza Sarancic

Supervisors:

Dr. Med Kjetil Retterstøl Dr. Med Trond Leren

Prof. Marit Veierød

Department of Nutrition, Faculty of Medicine UNIVERSITY OF OSLO

May 2012

(3)

ABSTRACT:

BACKGROUND AND AIMS: Familial hypercholesterolemia (FH) is an autosomal

dominantly inherited disorder of lipid metabolism, caused predominantly by mutations in the LDL receptor (LDLR) gene. Hypercholesterolemia, increased low-density lipoprotein

cholesterol (LDL-C) and increased total cholesterol are associated with an increased risk of cardio vascular diseases (CVD). Increased prevalence of CVD is associated with sudden premature death. Modern dietary and drug treatment have been associated with lower levels of LDL and total cholesterol. The primary aim of this study was to determine the mortality causes in the period 1992-2010, for the individuals that are diagnosed with FH. The secondary aim of this study was to investigate the prevalence of death by CVD among FH individuals with consideration of age, gender and death cause. The CVD death causes that will be under consideration are myocardial infarction (MI), cerebral infarction and aortic aneurysm.

STUDY DESIGN: The Medical Genetics Laboratory at Oslo University Hospital, has a FH registry. At endpoint the FH registry had 4688 individuals. The FH registry was linked to The Norwegian Cause of Death Registry. This linked registry contained 113 observed deaths. A cohort of men and women aged 0-59, were followed from 1992 to 2010. They were followed for 29853person years. Expected number of deaths was estimated by multiplying the gender, age and calendar specific deaths rates to the person years accumulated in the age and gender cohort. Deaths rates were estimated from the Norwegian population. Standardized mortality rate (SMR) was expressed as the ratio between observed deaths and number of expected deaths. Significance was defined by the 95% confidence interval (CI).

RESULTS: 46% of individuals with FH died of CVD. 30.1% died of cancer. The average age for death by CVD was 62.2 years. Women who died of CVD had an average death age of 67.2 years. For men average death age by CVD was 57.2 years. 71.2% of deaths by CVD were MI. SMR for death by CVD was 8.0 for the age group 30-39. For the age group 50-59, the SMR for death by CVD was 1.4. In the age 0-59 the SMR for death by CVD was

significant with 2.1.

CONCLUSION: At age 0-59, cancer and none-CVD death cause had lower total SMR among FH individuals, than the general population. Individuals with FH have higher prevalence of death by CVD, than the general population. The main death cause by CVD is MI. Women had a 10 years higher average age of death by CVD than men. At age 0-59, individuals with FH have a higher SMR of death by CVD, than the general population.

(4)

ACKNOWLEDGEMENTS

This master thesis presents a contribution to the knowledge on mortality outcomes to persons diagnosed with FH. The thesis has been conducted from August 2011 to May 2012 at the Department of Nutrition, Faculty of Medicine, University of Oslo.

I am very grateful for being given the opportunity to work with this master thesis.

I am thankful to dr.med Kjetil Retterstøl, for including me in this study. Big thanks to dr.med Kjetil Retterstøl for his comments, talks and scientific contributions. Without his supervisor teaching this thesis would not have been possible.

Thanks to my supervisors dr.med Trond Leren and prof. Marit Veierød for the scientific advices.

I would also like to thank the workers from the Molecular genetic laboratory, University Hospital in Oslo for making the FH registry. Thanks to the workers for the Norwegian public health institute for making The Norwegian Cause of Death Registry. Without these to

registries this master thesis would not have been possible.

Thanks to my family and friends for the encouragement during the study period. They have been a big inspiration for me.

Oslo, May 2012 Mirza Sarancic

(5)

List of Abbreviations

AAA Abdominal aortic aneurysm

ARH Autosomal recessive hypercholesterolemia

ATP Adenosine triphosphate Apo A-1 Apo lipoprotein A1 ApoB Apo lipoprotein B CHD Coronary heart disease CVD Cardiovascular disease CI Confidence interval EF Error factor

EGF-A Epidermal Growth Factor-Like Repeat A FH Familial hypercholesterolemia

HDL High- density lipoprotein

HDL-C High- density lipoprotein cholesterol HMG CoA Hydromethylglutaryl Co-enzyme A ICD International Classification of Diseases IDL Intermediated density lipoprotein IHD Ischaemic heart disease

LDL Low- density lipoprotein

LDL-C Low- density lipoprotein cholesterol LDLR Low-density lipoprotein receptor LDLRAP1 LDL receptor adaptor protein 1 LPL Lipoprotein lipase

(9)

MGL Medical Genetics Laboratory MI Myocardial infarction

PCSK 9 Proprotein convertase subtilisin/kexin type 9 PPAR Perixisome proliferator- activated receptor SES Socioeconomic status

SMR Standardized mortality rate SR-B Scavenger receptors class B

SREBP-2 Sterol regulatory element binding protein-2 SPSS Statistical Package of Social Sciences TAA Thoracic aortic aneurysm

TG Triglyceride

VLDL Very low- density lipoprotein VCAM-1 Vascular cell adhesion molecule-1 WHO World health organization

(10)

1. BACKGROUND

1.1 Hypercholesterolemia

Cholesterol can be synthesized de novo in the liver, gut and the central nervous system (1).

Cholesterol is essential to life. It is the primary component of cell membranes and a substrate for the synthesis of bile acids, steroid hormones and vitamin D (2).

Hypercholesterolemia is usually induced by a combination of genetic and lifestyle factors.

To avoid hypercholesterolemia it is important that the metabolism of cholesterol is in a normal biologically way (1).

Figure 1.0 Dietary cholesterol pathway (1).

Dietary cholesterol is absorbed in the gut. Chylomicrons are rich in triglycerides (TGs) and transport the dietary cholesterol to the live (3). In the liver cholesterol and TGs get modified. The liver secretes very low density lipoprotein (VLDL). Lipoprotein lipase (LPL) and hepatic lipase converts VLDL to intermediate density lipoprotein (IDL) and later to low density lipoprotein (LDL) (3, 4).

LDL distributes cholesterol to the body tissues. The uptake of LDL into the tissue cells is induced by LDLR (3).

(11)

Cells remove excessive cholesterol by high density lipoprotein (HDL). HDL transports excessive cholesterol back to the liver for degradation. Cholesterol is excreted through the bile as bile acids and free cholesterol (1). HDL can be taken up directly in the liver by

scavenger receptors class B (SR-B). Some of the HDL gets transferred to VLDL and LDL by the enzyme cholesteryl ester transfer protein. It is then taken up to the liver by LDLR (3).

Two major groups of lipoproteins are important in the cholesterol metabolism.

Apolipoprotein B (ApoB) is an important part of the chylomicrons, VLDL, IDL and LDL, while Apolipoprotein A1 (Apo A-1) is mainly found in HDL (3).

1.1.1 Inflammation and atherosclerosis

Atherosclerosis was characterized as artery formed lipid deposits (5). Now we have better knowledge about the atherosclerosis process and understand that inflammation plays a role(6).

Hypercholesterolemia promotes the inflammatory process. During the development of atherosclerosis leukocytes attach to the arterial wall by the vascular cell adhesion molecule- 1(VCAM-1) (4). Oxidized lipids can induce VCAM -1 expression. The macrophages express scavenger receptors for modified lipoproteins. This permits macrophages to ingest lipids and become foam cells in the arterial wall (4).

1.1.2 Risk factors for atherosclerosis and CVD Gender

Studies suggest that women have 0.25-0.3 mmol/l higher levels of HDL than men (7). It is assumed that the difference in levels of HDL, between the genders is caused by the estrogen hormone (7). A prospective study of 14786 Finnish individuals aged 25-64 years, suggested that the prevalence of coronary mortality was 5 times higher in men than in women (8).

Smoking

Smoking increases the risk for coronary heart disease (CHD) in both men and women. The risk of CHD increases with number of cigarettes (9). Smokers have lower levels of HDL cholesterol (HDL-C). Some studies suggest that smokers have 14% lower HDL levels, than none smokers (10).

(12)

Age

Aging is linked with increasing prevalence of atherosclerosis and CVD. 71% of Norwegian men, who died of CVD in 2006, were over 75 years old. 90% of Norwegian women, who died of CVD in 2006, were over 75 years old (11).

Diabetes

Type 2 diabetes mellitus is associated with a marked increase in the risk of atherosclerotic disease. Adults with type 2 diabetes mellitus are up to four times more likely than individuals without the disease, to suffer from cardiovascular events (12). It is assumed that insulin resistance can increase the prevalence of CVD by dyslipidemia, hypertension and inflammation (13).

Hypertension

Hypertension is a risk factor for the development of atherosclerosis. It is assumed that hypertension can induce inflammation on the arterial walls (14). Hypertension is associated with an increased risk of morbidity and mortality from cerebral infarction, CHD and MI (15).

There is often a prevalence of insulin resistance in hypertensive patients (16).

Abdominal obesity

Visceral fat is linked with higher levels of LDL and total cholesterol. Lower levels of HDL-C and increased prevalence of insulin resistance is also associated with visceral fat. Substances released by visceral fat, enter the portal vein and travel to the liver, where they can influence the production of blood lipids (17, 18). Visceral fat segregates free fatty acids, cytokines, tumor necrosis factor, interleukin-6 and adiponectin. High segregation of these substances can increase the risk of CVD by promoting insulin resistance and inflammation (17, 18).

Physical activity

Physical activity is important to reduce hypertension, diabetes and obesity (19). 5-13% of the risk for hypertension can be prevented by physical activity (20). Increased energy spender and utilization of fat are some of the health gains of physical activity. Physical activity can also decrease the prevalence of insulin resistance (19).

(13)

Stress

Stress can promote an unhealthy behavior like smoking, lack of exercise, excessive alcohol consumption and an unhealthy diet. Stressors can promote CVD through the segregation of hormones. Hormones like cortisol and adrenalin can affect the heart rhythm and the

contraction of blood vessels (21).

1.2 Familial hypercholesterolemia 1.2.1 Genetics

FH is an autosomal dominantly inherited lipid metabolism disorder (22).

Mutations in the LDLR-gene

Individuals with heterozygote FH have approximately 50% of the functioning LDLRs (23).

FH homozygotes can have 0-25% of normal LDLR function (24).

There has been identified over 1000 different mutations in the LDLR gene (25). Among individuals in Norway, approximately 130 LDLR gene mutations have been found to cause FH (22).

There are five main classes of mutations in the LDL-R (26). In class one none LDLRs are synthesized. LDLR gets processed in the Golgi apparatus (27). Disruption in the transport from the endoplasmic reticulum to the Golgi apparatus is categorized in class two. In class three mutations, LDL-R fails to bind to the LDL at the cell surface. Failing of clustering of the LDLR into the coated pits after binding to the LDL are class four mutations (28, 29).

Inside cells the ligand complex is transported to the endosomes, where the acidic environment causes dissociation of the receptor-ligand complex (27). The LDLR is recycled to the cell surface, while the LDL particle is degraded in the lysosomal compartment. Class five mutations prevent LDLRs from being recycled to the cell surface (28, 29).

Mutations in other genes than the LDLR-gene

Mutations in other genes than the LDLR gene can cause FH. Two of those gene mutations are ApoB and proprotein convertase subtilisin/kexin type 9 (PCSK9). Mutations in these genes are less common, than in the LDLR gene (27, 30).

(14)

ApoB is the major particle of the LDL. The LDLR binds to the LDL by ApoB (31).

PCSK9 induces degradation of the LDLR-protein by binding to the Epidermal Growth Factor- Like Repeat A domain (EGF-A) (31). Mutations in the gene coding for the PCSK9 occur rarely (27).

1.2.2 Frequency

FH is under diagnosed. It is estimated that only about 20% of the FH cases are ascertained.

Worldwide more than 10 million people have FH (22, 32).

The frequency of homozygote FH is one per 1000000 individuals. In most populations the frequency of heterozygote FH is approximately one per 500 individuals (22, 32). Some studies suggest that the Norwegian frequency of heterozygote FH is one per 313. It is estimated that approximately 15000 Norwegians have FH (33).

1.2.3 Diagnosis

FH is characterized by hypercholesterolemia, a total blood cholesterol level above 7.5mmol/l or a blood concentration of LDL-C above 4.9 mmol/l (34).

Table 1.0 Recommended blood levels of cholesterol for primary prevention of CVD.

From the Norwegian Directorate of Health (11, 35, 36).

Cholesterol Women Men

Total-cholesterol < 5.0 mmol/l < 5.0 mmol/l HDL-c > 1.3 mmol/l > 1.1 mmol/l LDL-c < 3.0 mmol/l < 3.0 mmol/l TG < 1.7 mmol/l < 1.7 mmol/l

To diagnose FH, hypercholesterolemia must be present together with xanthomatosis and a premature ischemic heart disease (IHD) history. The premature IHD history can be in the case or a close relative (34). Premature IHD history is by many defined as cardiovascular event occurring before 55 years among men and before 65 years among women (11).

Xanthomatosis is deposits of cholesterol in tendons and skin. The presence of xanthomas is correlated with a 3 times higher risk of CVD (37).

(15)

An alternative to the clinical criteria are DNA based analysis on LDLR, ApoB, and PCSK9 mutations (38).

1.2.4 Management of FH

Medical treatment

The mainly used drug treatments are statins. Statins provide reduction in plasma LDL by up regulating expression of LDLRs (39). Statins reduce the cholesterol production be inhibiting the enzyme 3-hydroksy-3-metylglutaryl-coenzym A (HMG-CoA)-reductase (40). HMG-CoA- reductase regulates the syntheses of cholesterol by controlling the transformation of HMG- CoA to mevalonic acid. This is the velocity determined step in the production of cholesterol (40). Sterol-regulatory element-binding protein 2 (SREBP-2) is a transcription factor, which gets activated when liver cells get deprived for cholesterol. SREBP-2 is a transcription factor for LDLR, HMG-CoA-reductase and 3-hydroxymethyl-3-glutaryl-(HMG-CoA) –synthase (40).

Studies suggest that statin prescription for FH individuals under 18 years can be beneficial in the prevention of atherosclerosis. Statin use did not indicate any adverse effect on growth and pubertal development. These studies prescribed a maximal dose of daily 40mg statin (41, 42, 43).

Ezetimibes and Bile acid sequestrants are mainly used when individuals experience statin intolerance or insufficient effect of statins.

Ezetimbe reduces the absorption of dietary and biliary cholesterol by preventing its

transport through the intestinal wall. Current knowledge indicates that Ezetimibe has no effect on the absorption of fat soluble vitamins, fatty acids and bile acids (44, 45).

Bile acid sequestrants are often called resins.Bile acid sequestrants interrupt the enterohepatic circulation of bile acids and inhibit the absorption of bile acids from the intestine. As result the liver cells produce more LDLRs. Bile acid sequestrants significantly reduce total cholesterol and LDL-C. Caution needs to be taken in therapy with bile acid sequestrants because these drugs may interfere with the absorption of fat soluble vitamins and concurrent medication (39, 46).

(16)

Fibrates and Nicotinic acids/niacin are sometimes used in the treatment of FH. Fibrates act on peroxisome proliferator activated receptors (PPARs). Activation of PPARs causes

transcription of a number of genes that facilitate lipid metabolism (46). Nicotinic acids /niacins have a modest effect on LDL-C. These drugs decrease the concentration of free fatty acids by inhibiting breakdown of adipose fat (46).

Dietary recommendations

Dietary advice should be given on addition to treatment with statins. Dietary changes should be combined with other lifestyle changes like daily exercise, smoking cessation and stress management (11, 34, 46).

Dietary intervention relies upon the reduction of LDL-C by replacing saturated fat with unsaturated fat. Some foods that contain unsaturated fats are nuts, fatty fish like salmon, mackerel and trout. Oils of raps, olive, soya and sunflower also contain unsaturated fatty acids. These foods should replace some of the saturated fat. Saturated fat is mainly found in bakery, diary and red meat products (11, 34, 46).

FH-individuals should be careful with consumption of cholesterol. Egg yolks are one of the biggest sources of cholesterol in food (11, 34, 46).

Plant sterols are assumed to reduce absorption of cholesterol. Plant sterols are mainly consumed as functional foods. It is also recommended with a nutrition that is rich in fiber.

Fibers are mainly found in wholegrain (11, 34, 46).

When it is necessary it is important to indicate a diet for weight loss (11, 34, 46).

(17)

Table 1.1 Nutrient composition of the recommended FH diet in Norway (47, 48).

Nutrient Recommended intake (of total energy intake) Total fat 25 %

Saturated fat < 7%

Trans fat < 1%

Monounsaturated fat 10-15%

Polyunsaturated fat 5-10%

- Omega-3 1%

Cholesterol < 200mg per day Protein 10-20%

Carbohydrat 50-60%

- Sugar Max 10%

Fiber 25-35 g per day Plant stanol/sterols 2 g per day Special treatments

Special treatments are mainly used by individuals with homozygous FH. LDL apheresis and liver transplantation are the most known special treatments (49).

LDL apheresis is a mechanical method, similar to that used in kidney dialysis. The process is removing LDL-C from the blood. LDL apheresis needs to be undertaken approximately every one to two weeks by trained health workers (49).

Liver transplantation gives individuals new functioning liver cells that are able to process the LDL-C. There are considerable side effects attached with liver transplantation (49).

Effects of treatment

Scandinavian Simvastatin Survival Study suggests that statins are reducing mortality and morbidity in CHD patients (50). Statins can lower LDL-C by an average of 1.8mmol/l.This reduces the risk of IHD events by 60% and stroke by 17% (51).

Compared to alone statin treatment, combination of statin with bile acid sequestrants can reduce LDL-C by additionally 10-20% (52). Ezetimibe has largely replaced the use of bile acid sequestrants. When used as monotherapy, ezetimibe decreases LDL-C by about 18% (44, 45). Ezetimibe combined with any statin will provide a mean LDL-C reduction of 14 to 25%

(44).

(18)

Several studies suggest that dietary recommendations can have impressive impact on lowering LDL-C. The mainly used polyunsaturated fatty acid replacements for saturated fatty acids are ω – 6. Some studies suggest that replacing 5% of the saturated fatty acids with polyunsaturated fatty acids decreases the ratio between total and HDL cholesterol by 0.17 (53).

A meta-analysis suggested that a daily intake of stanol /sterol enriched fat spreads

significantly reduced total cholesterol by 7-11%. The study used stanol/sterol dosage of 2.3 g per day (54).

Recommended dietary fiber intake is suggested to reduce total cholesterol. In

hypercholesterolemic and diabetic patients, soluble fibers with diet that is low at saturated fat and cholesterol, lowers LDL-C with 5-10 % (55, 56).

1.3 Methodological problems in FH mortality

For ethical reasons placebo controlled studies that measure mortality and hard endpoint, can not be conducted among individuals with FH (57). Little is known how long term statin treatment has affected morality among individuals with FH. An example of this is that in clinical trials statin treatment has been studied for no more than two years in FH children (58, 59). New data are needed for answering questions about how modern treatment affects the mortality prognosis.

Some individuals get FH diagnosed late in life. In early life years, these individuals are exposed for high cholesterol levels. It is unclear how high cholesterol levels before the introduction of treatment, has influenced the prevalence of death by CVD.

Previous studies

The time frame of the Simon Broome study in the early 1990s overlaps with the introduction of statins. This allows a comparison of SMR before and after the use of these medications.

The relative risk of coronary mortality in patients aged 20–59 years was higher from 1980 to 1991, than from 1992. Simon Broome study suggested that statin treatment is effective in lowering the risk of death from CHD in individuals with FH (60).

(19)

2146 FH patients were in the Rotterdam study. FH patients received a mean daily dose of 33mg simvastatin or 49mg atorvastatin. A compared group did not receive treatment. In the Rotterdam study the risk of MI in the statin treated patients was not significantly greater than in an age matched sample from the general population. In the treated group, the Rotterdam study observed an overall risk reduction of MI by 76% (61).

1.3.1 Challenges in assessment of FH mortality

. Individuals who avoid treatment

Some FH individuals can not get treatment in some periods of life. It is unknown how these periods affect mortality and hard endpoints.

It is recommended for women during pregnancy to avoid statins (62). Statins have been linked to fetal abnormalities. It is assumed that first-trimester statin exposure can defect the central nervous system and make unilateral limb deficiencies. Ezetimbie and Niacin have in animal studies adverse effects on the fetus (63).

Some studies suggest that muscular side effects during exercise are related to statin treatment (64). A study monitored 22 professional athletes with FH for 8 years. The study suggested that only 20% of the professional athletes tolerated statin treatment without getting muscular side effects (65).

It is unclear how long term use of statin and other lipid lowering drugs can influence the development of malignant diseases. Most studies suggest that long term statin use is not associated with cancer (66, 67, 68, 69).

Gender

The Simon Broom study suggested that women aged 20-39, who were treated for FH, had an annual coronary mortality of 0.17%. For men in the same category the annual coronary mortality was 0.46%. For FH men and FH women aged 60-79 the annual mortality was 1.1%

(60).

Age

A Japanese study observed 527 individuals with heterozygote FH. Individuals were examined over 10 years. 41 deaths were observed. The mean age of death from a cardiac event was

(20)

significantly younger among men. Men suffered from a cardiac event at average of 54 years, while women suffered from a cardiac event at average of 68 years (70, 71).

Differences in national nutrition policy

CVD is responsible of about 50% of all deaths in Europe. CVD is the main contributor to the 20 year difference, in life expectancy across Europe (72).

One example of how national nutrition policy can affect prevalence of CVD mortality is the North Karelia project. The North Karelia project brought dietary changes to a local population in Finland (72). The project primary targeted changes in smoking and dietary habits. The results of the North Karelia project suggested that reducing risk factors can have a dramatic effect on CHD mortality.In 35 years, the annual age adjusted CHD mortality rate among the 35- 64 year old male population in North Karelia declined with 85% (73).

(21)

1.4 FH and death causes

It has been estimated that 200000 individuals with FH, die of CHD each year (22). Increased total cholesterol levels are associated with CVD. Increased prevalence of CVD is associated with increased sudden premature death (74).

Figure 1.1 Increased cholesterol increase the risk of death by MI and IHD (74 , 75).

Logarithmical graph describes the rate of death by MI and other IHD with consideration of cholesterol levels. Deaths are calculated as rate per 100000 per year (75).

It is estimated that 50% of men that are untreated for heterozygote FH, will get CVD before 50 years of age. Before 60 years of age 50% of women that are untreated for

heterozygote FH, will get CVD (76). The life age for both genders is estimated to be reduced between 20-30 years (77).

1.4.1 FH and myocardial infarction (MI)

Studies before the statin treatment decade suggest that individuals with untreated FH are 11 times at greater risk for CHD death than the general population (70). Before the introduction of statins, some studies suggested that about 70% of deaths among FH individuals were from CHD (70, 71). Some studies suggest that since the introduction of statin treatment, CHD mortality has significantly reduced by 37% among FH individuals (78). Other studies suggest

(22)

that the risk of MI in statin treated patients was not significantly greater than in an age matched sample from the general population (61). In 1994 among the Norwegian population, 20.5% of women and 26.8% of men died of MI (79).

1.4.2 FH and cerebral infarction

Kaste et al conducted a study before the treatment with statins was introduced. They suggested that individuals with FH have 20 times higher risk of cerebral infarction than the general population (80, 81). Other studies suggest that FH individuals with statin treatment are not at higher risk of fatal stroke than the general population (82). Cerebral infarction is the third most prevalent death cause in Norway. In 1999, 12% of the Norwegian population died from cerebral infarction (79).

1.4.3 FH and Aortic aneurysm

In western countries ruptured abdominal aortic aneurysm (AAA) is the cause of 1-2 % of all deaths (83). Men, smokers and individuals with increasing age are at higher risk for AAA, but increased cholesterol levels, hypertension and low HDL levels also increases the prevalence of AAA (84).

(23)

1.5 Aims of the study

1.5.1 Problems to address regarding FH and mortality

High LDL and total cholesterol levels are associated with increased risk of CVD (74, 75).

High prevalence of CVD is associated with increased premature mortality (85). Individuals with FH are at risk for high LDL and total cholesterol. Modern dietary interventions and drug treatments have been associated with lower levels of LDL and total cholesterol (11, 51). We have little knowledge about how modern treatment has affected the CVD morality prognosis.

It can be assumed that premature mortality among individuals that have been diagnosed with FH has been reduced the last years. There is need for more knowledge about the relationship between FH and mortality.

1.5.2 Aims

The aims of this study were to investigate:

• The mortality causes in the period 1992-2010, for the individuals who are diagnosed with FH.

• The prevalence of death by CVD among FH individuals with consideration of age, gender and death cause. The CVD death causes that will be under consideration are MI, cerebral infarction and aortic aneurysm.

(24)

2. SUBJECTS AND METHODS

2.1 Study design and data collection

The data was derived from patient lists at the Medical Genetics Laboratory (MGL) and The Norwegian Cause of Death Registry.

Registries

Molecular genetic testing for FH has been available in Norway since 1998 at a national MGL in Oslo University Hospital. Per December 2010 there were registered over 23000 patients of which 4688 with diagnosis of FH.

The Norwegian Cause of Death Registry is a population based registry of all causes of death since 1951 (86). Doctors are required to complete a death certificate of all reported deaths. Death certificates are collected by the Cause of Death Registry. The coding system used in the death certificates is the International Classification of Diseases (ICD). ICD is used by the world health organization (WHO) (87). This international system allows us to follow development of various causes of death and to compare mortality causes between different countries. Currently the 10th revision of ICD is being used.

Norwegian Cause of Death Registry and the FH registry were linked by the governmental statistical bureau.

Table 2.0 Variables from the linked registry

Categorical variables Continuous variables Other variables

 Gender

 The death cause

 The reliability of the diagnosis

 Diagnose 2

 Diagnose 3

 Diagnose 4

 Diagnose 5

 The place of mortality

 Death in institution

 The commune were the case lived

 Datum of mortality

 Year of mortality

 Age of death

 The circumstances of death 1

 The circumstances of death 2

(25)

2.2 Study population

At the study endpoint, the FH registry contained 4688 individuals. Of the 4688 individuals 2238 were men and 2450 were women. The linked registry contained 113 observed deaths.

The birth year of the individuals in the FH registry, suggest that the FH registry has an approximate normal distribution of age.

The average age in the FH registry was 41.6 years. The youngest person in the FH registry was 110 days old. The oldest person in the FH registry was 94.7 years. Five persons were under one year of age, while six persons were over 90 years of age.

Figure 2.0 Distribution of birth year in the FH registry.

(26)

2.3 Categorization of the death causes

Three cases were coded with ICD-9. Rest of the cases was coded with ICD-10. We categorized death causes into the following categories:

1. Death by CVD 2. Death by cancer 3. Death by other causes 4. Death by possible CVD

CVD death causes were categorized into following categories:

1. Death by MI

2. Death by cerebral infarction 3. Death by Aortic aneurysm 4. Death by possible MI

(27)

Table 2.1 ICD-codes that were interpreted as death by CVD. Categorized CVD death causes (87, 88, 89).

ICD codes Description CVD categorization I21.9 Acute myocardial infarction, unspecified MI

I22.9 Subsequent myocardial infarction of unspecified site MI I24.8 Other forms of acute ischemic heart disease MI

I25.1 Atherosclerotic heart disease MI I25.2 Old myocardial infarction MI

I25.8 Other forms of chronic ischaemic heart disease MI I25.9 Chronic ischaemic heart disease, unspecified MI I26.9

I35.0 I35.1 I38 I48 I50.1 I51.7 I51.9 I60.9 I62.0 I63.9 I64 I71.1 I73.9 I80.2 E78.0 410(ICD-9)

Pulmonary embolism without mention of Possible MI acute cor pulmonale

Aortic (valve) stenosis Aortic aneurysm Aortic (valve) insufficiency Aortic aneurysm Endocarditis, valve unspecified Possible MI Atrial fibrillation and flutter Possible MI Left ventricular failure MI

Cardiomegaly Possible MI Heart disease, unspecified MI

Subarachnoid haemorrhage, unspecified Cerebral infarction Subdural haemorrhage (acute) (nontraumatic) Cerebral infarction Cerebral infarction, unspecified. Cerebral infarction Stroke, not specified as haemorrhage or infarction Cerebral infarction Thoracic aortic aneurysm, ruptured Aortic aneurysm Peripheral vascular disease, MI

unspecified Claudicatio intermittens

Phlebitis and thrombophlebitis of Possible MI other deep vessels of lower extremities

Pure hypercholesterolaemia MI Acute myocardial infarction MI

(28)

Table 2.2 ICD-codes that were interpreted as death by cancer (87, 88, 89).

ICD codes Description

C06.9 Mouth, unspecified

C18.0 Caecum

C18.9 Colon, unspecified

C20 Malignant neoplasm of rectum

C22.1 Intrahepatic bile duct carcinoma

C26.9 Ill-defined sites within the digestive system C34.1

C34.9

Upper lobe, bronchus or lung Bronchus or lung, unspecified C43.6

C43.9 C50.9 C53.9 C56 C61 C71.2 C74.9 C80 C91.1 147(ICD-9) 191(ICD-9)

Malignant melanoma of upper limb, including shoulder Malignant melanoma of skin, unspecified

Breast, unspecified Cervix uteri, unspecified Malignant neoplasm of ovary Malignant neoplasm of prostate Temporal lobe

Adrenal gland, unspecified

Malignant neoplasm, without specification of site Chronic lymphocytic leukaemia of B-cell type Pharynx

Malignant neoplasm of brain

(29)

Table 2.3 ICD-codes that were interpreted as death by other causes (87, 88, 89).

ICD codes Description A 40.1

A41.9 G30.9 G31.9 G40.9 G71.3 K26.4 K57.8

K81.9 R54 R99.9

V23.4 V48.5

W15 X42

X44

X59.0 X61

Y08

Sepsis due to streptococcus, group B Sepsis, unspecified

Alzheimer's disease, unspecified

Degenerative disease of nervous system, unspecified Epilepsy, unspecified

Mitochondrial myopathy, not elsewhere classified Duodenal ulcer

Diverticular disease of intestine, part unspecified, with perforation and abscess

Cholecystitis, unspecified Senility

Other ill-defined and unspecified causes of mortality

Motorcycle rider injured in collision with car, pick-up truck or van

Car occupant injured in noncollision transport accident Fall from cliff

Accidental poisoning by and exposure to narcotics and psychodysleptics [hallucinogens], not elsewhere classified

Accidental poisoning by and exposure to other and unspecified drugs, medicaments and biological substances

Exposure to unspecified factor causing fracture Intentional self-poisoning by and exposure to antiepileptic, sedative- hypnotic, antiparkinsonism and psychotropic drugs, not elsewhere classified

Assault by other specified means

(30)

Table 2.4 ICD-codes that were interpreted as death by possible CVD (87, 88, 89).

ICD codes Description

E10.9 E14.9 W10.0

Insulin-dependent diabetes mellitus without complications Unspecified diabetes mellitus without complications

Fall on and from stairs and steps

2.4 Statistics

The analyses were performed using The Statistical Package of Social Sciences (SPSS) version 19.0. Some of the analysis was performed by using Microsoft Excel.

Descriptive statistics was used to describe the frequency of the gender and age.

Person years are the estimation between when a person was included in the FH registry and when that person reached endpoint.

The age and calendar specific deaths rates for men and women in the Norwegian population were found from the governmental statistical bureau. Expected number of deaths was

estimated by multiplying the gender, age and calendar specific deaths rates to the person years accumulated in the age and gender cohort (60, 90).

SMR was derived from the ratio of the number of observed deaths, to the number of expected deaths (60, 90). SMR was calculated by indirect standardization.

Absolute risks of mortality were calculated per 100000 person years.

95% CI was used to determine statistical significance. 95 % CI was calculated with the mathematical equation 95% CI= SMR/EF to SMR x EF where EF stands for error factor. EF=

exponential (1.96/ √observed number of deaths). If 95% CI contained number 1, then the SMR was considered as none significant (90). Significant 95% CI is assumed to give a significance with a p-value <0.05 (90).

2.5 Ethics

The study was approved by The Regional Ethic Committee (appendix 1) and by Oslo University Hospital for internal control and settling of research responsibility (appendix 2).

(31)

3. RESULTS

3.1 Description of total mortality

The linked registry contained 113 observed deaths. 46.0% of observed deaths were from CVD. The next highest death cause was cancer with 30.1%. Deaths by other causes like accidents and unspecified sepsis represented 20.4% of the observed deaths. Death by possible CVD was responsible for 2.7% of observed deaths.

Table 3.0 Observed deaths.

Observed deaths Percent

Death by CVD 52 46.0

Death by cancer 34 30.1

Death by other causes 24 21.2 Death by possible CVD 3 2.7

Total 113 100.0

The age of death among the 113 observed deaths were wide spreading. The youngest death of age was 18 years, while the oldest was 94 years. This represents a life year difference in 76 years. The average age of death was 61.1 years.

Figure 3.0 Age of death among all death causes.

(32)

The FH population had a peak for excepted death around 60 years. The general Norwegian population had a peak for expected death around 80 years. The prevalence of expected death was higher among the FH population than the general Norwegian population in the age 0-60.

In the age after 60 the prevalence of expected death was higher among the general Norwegian population, than the FH population.

Figure 3.1 Expected deaths among FH and the general Norwegian population. This figure compares expected death for 4666 persons from the FH registry with 4666 persons from the general Norwegian population. Expected deaths for the general Norwegian population were calculated as described in section 2.4. Expected deaths for 10 years age groups were divided with total expected deaths. Numbers from each age group were multiplied with 4666. The same estimates were calculated for the FH population. For the FH population the estimates were calculated from observed deaths.

(33)

3.2 Description of total mortality with consideration of gender and age

The linked registry had a composition of 59 dead men and 54 dead women. This indicates an equal gender composition.

Table 3.1 Observed deaths among genders.

Gender Observed deaths Percent

Men 59 52.2

Women 54 47.8

Both genders 113 100.0

For men the average age of death was 58.1 years. The lowest death age among men was 18 years. The highest death age among men was 85 years.

For women the average age of death was 64.4 years. The lowest death age among women was 31 years, while the highest death age among women was 94 years.

Table 3.2 Age of death among genders.

Mean Minimum Maximum Lower 95% CI Upper 95% CI

Men 58.1 18 85 53.7 62.5

Women 64.4 31 94 60.0 68.9

Both genders 61.1 18 94 58.0 64.3

Figure 3.2 Age of death among men.

(34)

Figure 3.3 Age of death among women.

All the cases that died before 30 years of age were men. None of these cases died of CVD.

Table 3.21 Cases that died before 30 years of age.

ICD-code Description Age of death Gender

X61 Intentional self-poisoning by and exposure to antiepileptic, sedative-hypnotic,

antiparkinsonism and psychotropic drugs, not elsewhere classified

20 Men

X42 Accidental poisoning by and exposure to narcotics and psychodysleptics [hallucinogens], not elsewhere classified

24 Men

W15 Fall from cliff 19 Men

V48.5 Car occupant injured in noncollision transport accident

19 Men

V23.4 Motorcycle rider injured in collision with car, pick-up truck or van

18 Men

All the cases that died after 85 years of age were women. There were five cases that died after 85 years of age. Of these four persons died of CVD.

(35)

Table 3.22 Cases that died after 85 years of age.

ICD-code Description Age of death Gender I48 Atrial fibrillation and flutter 89 Women I35.0 Aortic (valve) stenosis 90 Women I63.9 Cerebral infarction, unspecified 91 Women

R54 Senility 94 Women

I38 Endocarditis, valve unspecified 87 Women 3.3 Analysis of none CVD mortality

None CVD mortality

None CVD mortality had 26 observed deaths for age 0-59. In the age group 10-19 did FH individuals have higher SMR, than the general population for none CVD death. Total SMR for none CVD mortality for age 0-59 was significant with 0.6.

Table 3.3 Mortality analysis: None-CVD mortality both genders (All ICD-10 codes besides I00-99).

Attained age

Person- years of observation

Observed deaths

Expected Deaths

SMR 95% CI Absolute risk per 100000 person years

0-9 461 0 0.2 0.0 0 0

10-19 3093 3 0.8 3.8* 1.2-11.8 97

20-29 5726 2 3.6 0.6 0.2-2.4 35

30-39 6265 3 4.8 0.6 0.2-1.9 48

40-49 7533 6 11.0 0.5 0.2-1.1 80

50-59 6775 12 23.1 0.5* 0.3-0.9 177

Total 29853 26 43.5 0.6* 0.4-0.9 87

* p<0.05

Cancer

Cancer had 11 observed deaths in the age 0-59. In the age 0-29, there were none observed deaths by cancer. In the age group 30-39 did FH individuals have higher SMR for deaths by cancer, than the general population. Total SMR for cancer mortality at age 0-59 was 0.6.

(36)

Table 3.31 Mortality analysis: Cancer both genders (ICD-10 codes C00-97 and D00-47).

0-9 461 0 0.0 0.0 0 0

10-19 3093 0 0.1 0.0 0 0

20-29 5726 0 0.3 0.0 0 0

30-39 6265 2 1.1 1.8 0.5-7.2 32

40-49 7533 2 4.7 0.4 0.1-1.6 27

50-59 6775 7 13.3 0.5 0.2-1.1 103

Total 29853 11 19.5 0.6 0.3-1.1 37

* p<0.05

3.4 Analysis of none CVD mortality with consideration of gender and age None CVD mortality

None observed CVD deaths were among women, in the age group 0-19. It was 13 observed deaths for none CVD mortality among women. For men it was observed 13 none CVD deaths. Men in the age group 0-19 and women in the age group 20-39 had higher SMR for none CVD death, than the general population. Total SMR for none CVD mortality among men in the age 0-59 was significant with 0.5. Total SMR for none CVD mortality among women in the age 0-59 was 0.7.

Table 3.4 Mortality analysis: None-CVD with consideration of genders (All ICD-10 codes besides I00-I99).

Observed Deaths

SMR 95% CI Absolute risk per 100000 person years Men

0-19 1946 3 0.9 3.3* 1.1-10.2 154

20-39 5806 2 5.6 0.4 0.1-1.6 34

40-59 6980 8 18.7 0.4* 0.2-0.8 115

Total 14732 13 25.2 0.5* 0.3-0.9 88

Women

0-19 1608 0 0.5 0.0 0 0

20-39 6185 3 2.6 1.2 0.4-3.7 49

40-59 7328 10 14.8 0.7 0.4-1.3 136

Total 15121 13 17.9 0.7 0.4-1.2 86

* p<0.05

(37)

Cancer

In the age 0-59 there was 3 observed deaths among men and 8 among women for death by cancer. In the age 0-39, there were none observed deaths by cancer among men. The mortality analysis with consideration of gender suggests that women in the age group 20-39 have higher SMR, than the general population for death by cancer. For death by cancer men in the age 0- 59 had a significant total SMR with 0.3. For women in the same age group SMR for death by cancer was 0.8.

Table 3.41 Mortality analysis: Cancer with consideration of gender (ICD-10 codes C00-97 and D00-47).

SMR 95% CI Absolute risk per 100000 person years Men

0-19 1946 0 0.1 0.0 0 0

20-39 5806 0 0.6 0.0 0 0

40-59 6980 3 8.0 0.4 0.1-1.2 43

Total 14732 3 8.7 0.3* 0.1-0.9 20

Women

0-19 1608 0 0.1 0.0 0 0

20-39 6185 2 0.8 2.5 0.6-10.0 32

40-59 7328 6 9.6 0.6 0.3-1.3 82

Total 15121 8 10.5 0.8 0.4-1.6 53

* p<0.05

(38)

3.5 Description of CVD mortality

52 persons died of CVD. The average age of death by CVD was 62.2 years. Most persons died of CVD in the age group 60-69. It was 12 CVD deaths in the age 30-49, while it was 7 CVD deaths in the age 80-99.

Figure 3.4 Age of death among CVD deaths.

The FH population had a peak of around 60 years for expected CVD death. The general Norwegian population had a peak of around 80 years for expected CVD deaths. The prevalence of expected CVD deaths was higher among the FH population than the general Norwegian population in the approximant age 25-60. In the age after 60 the prevalence of expected CVD deaths was higher among the general Norwegian population compared to the FH population.

(39)

Figure 3.5 Expected CVD deaths among FH and the general Norwegian population. This figure compares expected CVD death for 4666 persons from the FH registry with 4666 persons from the general Norwegian population. The figure was calculated by the same method as figure 3.1.

3.6 Description of CVD mortality with consideration of gender and age.

Both men and women had 26 observed CVD deaths.

Table 3.5 Observed CVD deaths among genders.

Observed CVD deaths Percent

Men 26 50.0

Women 26 50.0

Both genders 52 100.0

37 observed deaths were from MI. It represented 71.2% of total CVD mortality. Six persons died of cerebral infarction and three died of aortic aneurysm.

(40)

Table 3.6 Categorized CVD deaths among both genders.

Death causes Observed deaths Percent Both genders

Death cause by MI 37 71.2

Death by cerebral infarction 6 11.5

Death by Aortic aneurysm 3 5.8

Possible MI 6 11.5

Total 52 100.0

84.6% of CVD deaths among men were from MI. There were 22 observed MI deaths among men. One man died of cerebral infarction while two men died of aortic aneurysm.

Table 3.61 Categorized CVD deaths among men.

Death causes Observed deaths Percent Men

Death by cerebral infarction 1 3.9 Death by Aortic aneurysm 2 7.7

Total 26 100.0

15 women died of MI. This represented 57.7% of CVD mortality among women. One woman died of aortic aneurysm. 19.2% of CVD mortality among women was from cerebral

infarction. 5 women died of cerebral infarction.

Table 3.62 Categorized CVD deaths among women.

Death causes Observed deaths Percent Women

Death by cerebral infarction 5 19.2

Death by Aortic aneurysm 1 3.9

Total 26 100.0

Women had an average age of death by CVD at 67.2 years. Men had an average age of death by CVD at 57.2 years. Women had higher age of death, than men in all the CVD death categories. Men who died of MI had an average age of death at 56.6 years. Women who died of MI had an average age of death at 62.8 years. For both genders the average age of death by MI was 59.0 years. Persons who died of aortic aneurysm had the highest average death age with 76.3 years.

(41)

Table 3.63 Average age of death among categorized CVD deaths.

Death causes Men Women Both genders Death cause by MI 56.4 62.8 59.0

Death by cerebral infarction 61.0 68.4 67.2 Death by Aortic aneurysm 69.5 90.0 76.3

Possible MI 47.0 74.8 70.2

Total 57.2 67.2 62.2

Figure 3.6 Age of CVD deaths among genders

(42)

3.7 Analysis of CVD mortality

Age 0-29 did not have observed CVD deaths. Age 30-59 had higher SMR, than the general population for death by CVD. SMR for death by CVD decreases with increasing age for the age 30-59. Total SMR for death by CVD was significant with 2.1.

Table 3.7 Mortality analysis: CVD both genders (ICD-10 codes I00-I99).

0-9 461 0 0.0 0 0 0

10-19 3093 0 0.0 0 0 0

20-29 5726 0 0.1 0 0 0

30-39 6265 4 0.5 8.0* 3.0-21.3 64

40-49 7533 8 2.5 3.2* 1.6-6.4 106

50-59 6775 10 7.2 1.4 0.8-2.6 147

Total 29853 22 10.3 2.1* 1.4-3.2 74

*p<0.05

3.8 Analysis of CVD mortality with consideration of gender and age

The mortality analysis of death by CVD with consideration of gender suggests that both genders have a higher SMR than the general population in age 0-59. SMR is decreasing with increasing age for both genders after 20 years of age. For men the total SMR for CVD death in the age 0-59 was 1.7. For women the total SMR for CVD death in the age 0-59 was significant with 3.6.

Table 3.8 Mortality analysis: CVD with consideration of gender (ICD-10 codes I00-I99)

SMR 95% CI Absolute risk per 100000 person years Men

0-19 1946 0 0.0 0.0 0 0

20-39 5806 2 0.5 4.0 1.0-16.0 34

40-59 6980 11 7.1 1.6 0.9-2.9 158

Total 14732 13 7.6 1.7 1.0-2.9 88

Women

0-19 1608 0 0.0 0.0 0 0

20-39 6185 2 0.2 10.0* 2.5-40.0 32

40-59 7328 7 2.3 3.0* 1.4-6.3 96

Total 15121 9 2.5 3.6* 1.9-6.9 60

*p<0.05

(43)

3.9 Analysis of mortality by CVD with consideration of periodic intervals

In the age 0-59 both periods had a significant higher SMR than the general populations for death by CVD. SMR is decreasing with increasing age for both genders after 20 years of age.

In period 1992-2005 the total SMR for CVD death in the age 0-59 was 1.9.

Table 3.9 Mortality analysis: CVD 1992-2005 both genders (ICD-10 codes I00-I99).

0-19 928 0 0.0 0.0 0 0

20-39 5958 3 0.4 7.5* 2.4-23.2 50

40-59 7446 8 5.5 1.5 0.8-3.0 107

Total 14332 11 5.9 1.9* 1.1-3.4 77

*p<0.05

In period 2006-2010 the total SMR for CVD death in the age 0-59 was significant with 3.0.

Table 3.91 Mortality analysis: CVD 2006-2010 both genders (ICD-10 codes I00-I99).

0-19 2626 0 0.0 0.0 0.0 0

20-39 6030 1 0.3 3.3 0.5-23.4 17

40-59 6866 10 3.4 2.9* 1.6-5.4 146

Total 15522 11 3.7 3.0* 1.7-5.4 71

*p<0.05

(44)

4. DISCUSSION

4.1 Challenges in the study

Challenges in the categorization of the death causes

In the Norwegian somatic health care, ICD-10 was officially in use 1. January 1999. The WHO introduced ICD-10 in 1993 (91). In our study most of the cases from the period 1992 to 1999 were coded with ICD-10. Three of the deaths were coded with ICD-9. This could

suggest a difference in the criteria of death coding between our study population and the general population.

According to the ICD-10 manual, all diseases in the circulatory system are categorized with category block I (87, 88, 89). Persons that had the category block I in the main diagnosis, were defined as death by CVD. Five persons with death cause E78.0 pure

hypercholesterolemia were also defined as death by CVD. ICD-10 code E78.0 also includes FH (87, 88, 89). Many studies suggest that individuals with FH are at higher relative risk of death by CVD (70, 71).) This was the reason why we assumed that persons with E78.0 died of MI.

Three persons were defined as possible death by CVD. One person had main death cause insulin dependent diabetes mellitus without complications. The other had main death cause, unspecified diabetes mellitus without complications.Type 2 diabetes mellitus is associated with increased risk of CVD (12, 13). The third possible case had main death cause as fall on and from stairs and steps. This case had several other diagnoses. These other diagnosis included instantaneous death, unspecified CVD and unspecified injury. We can assume that this person has had a cardiovascular event, before the fall from the stairs. It is likely that these three cases died of CVD, but it is impossible to claim with scientific knowledge that these cases died of CVD.

Challenges in the categorization of CVD death causes

Some of the ICD-codes that described the death cause were difficult to categorize in the different CVD categorizes. One case died of peripheral vascular disease, unspecified claudicatio intermittens. 60% of individuals with peripheral vascular disease, unspecified claudicatio intermittens die of MI (93, 94). We assumed that this person died of MI, since

(45)

claudicatio intermittens affects the legs and death can only occur as a complication of claudicatio intermittens.

The person that died of left ventricular failure was defined as death by MI. Left ventricular failure is a usual symptom of MI (94, 95).

Some death causes were interpreted as possible MI. 2 persons died of unspecified endocarditis valve. According to the Norwegian doctors hand manual is heart failure the mostly common symptom for endocarditis (96). Cerebral embolism is prevalent in 30% of the endocarditis cases (97).

Studies suggest that atrial fibrillation and flutter are important risk factors for myocardial and cerebral infarctions (98, 99). Atrial fibrillation and flutter is assumed to increase the risk of cerebral infarction (100). Some studies suggest that atrial fibrillation and flutter increases the risk of sudden cardiac death by a 1.31 risk ratio (101). It can be interpreted that the case that died of atrial fibrillation and flutter is at equal risk for myocardial and cerebral infarction.

One person with death cause I80.2 had phlebitis and thrombophlebitis of other deep vessels of lower extremities. Major risk factor for pulmonary infarction is deep vein thrombosis in the leg, which emigrates to the lung arteries. This can cause sudden death by preventing blood and oxygen supplement to the lungs (102). One person died of pulmonary embolism without mention of acute cor pulmonale. We assume that these two cases died of pulmonary infarction.

One person died of cardiomegaly. Cardiomegaly is associated with abnormal heart rhythms, heart valve problems, atherosclerotic disease and sometimes chronic diseases such as thyroid disorders (103, 104). Some of the complications of having enlarged heart are sudden death, heart failure and cardiac arrest (104). It is difficult to determine if this person died of MI.

Challenges in the control of risk factors

Many of the factors that could influence the prevalence of death by CVD are not excluded. 15 of the persons who died of CVD had other diagnosis.

Mostly the other diagnosis confirmed the CVD death cause. Other diagnoses that probably influence the prevalence of death by CVD like hypertension, obesity and diabetes mellitus should have been diagnosed as other diagnosis (13, 14, 18).

(46)

The diagnosis unspecified diabetes mellitus without complications is associated with a marked increased risk in the development of atherosclerotic disease (12, 13). One person died of chronic ischaemic heart disease at 66 years of age. This person had also diagnosis of unspecified diabetes mellitus without complications. At age 54, one person died of old myocardial infarction. This person had several other diagnoses. One of the diagnoses was unspecified diabetes mellitus without complications. We can assume that these two cases would have had longer life age, if they did not have unspecified diabetes mellitus without complications.

Dementia often disables individuals to function in their daily life by reducing cognitive functions. Some individuals with dementia are not able to have a normal memory, attention and problem solving skills (105, 106). One person died of unspecified chronic ischaemic heart disease. This person had secondary diagnosis unspecified dementia. We can assume that this person has forgotten to have a good adherence to dietary and medical recommendations. If this is correct then we can assume that this person would have had a higher life age.

According to WHO, among patients with chronic illness, approximately 50% do not take medications as prescribed (107). Other long term studies suggest that adherence to statin therapy is low (108). A study analyzed a database with 21393 subjects who received at least one prescription for statins during the period between 1994 and 2003. The adherence to statin therapy reminded low with a with a 50% discontinuation rate in the first year (108).

A study with 336 FH patients examined adherence to the use of medications. A total of 36.6% reported total adherence to the use of cholesterol lowering medication. 64.4% reported some level of none adherence, but only 15% reported that they had quit medical use. The study suggested that patients with FH had overall high levels of adherence to the use of medication (109).

It is unknown if the study population had good adherence to the dietary and medical recommendations. Adherence to dietary and drug advices could have affected the mortality age.

(47)

4.2 Methodological consideration

Methodological consideration of person years

We did not have the date when the first 704 persons were included in the FH registry. We estimated for all besides of one case, that they were included in the FH registry in 01.01.1992.

This case was born after 01.01.1992. We estimated that this case was included in the FH registry the day after birth. The person years would probably have had other estimates, if we had the accurate datum of the registration of all cases.

This study did not take consideration of censored cases. It can be assumed that some persons have changed their ID-number or that they have emigrated to other countries. It is possible that some persons could have been excluded from the FH registry, before our end point. In the Simon Broome study five cases were censored because of emigration (60). Our study had about four times more individuals than the Simon Brome study. If we can expect to have four times more censored cases than the Simon Broome study, then this study would have had around 20 censored cases. The estimate of person years would have been lower, if the study had taken consideration of censored cases, but still 20 of 4688 is a small fraction.

Methodological limitations in the statistical method

Our study used 95% CI to test if the results were significant. Many other studies use p-value to test significance. Significance by 95% CI and the p-value are estimated from the same test observation. This makes a similarity between the 95% CI and p-value. Usually a significant 95% CI is assumed to give a significance with a p-value <0.05 (90). It is therefore not possible to claim that significant results defined by the 95% CI are more unstable than significant result defined by the p-value (90).

Some of the results from the mortality analysis were not significant. The study samples were probably too small to give significant results for every age group (90).

Standardization facilitates comparisons between populations with distinct gender and age structures (110). The most used statistical methods for standardization are direct

standardization and indirect standardization. Each of these standardization methods has limitations. The direct standardization indicates instability, when the populations to be age adjusted are small. Even for small age adjusted populations indirect standardization achieves

(48)

precision in the estimates. It is difficult to conduct a valid internal comparison in the study population with indirect standardization (110, 111).

For both standardization methods the main problem is selecting a proper population that is multiplied with the person years (111).

Individuals with FH are in this study compared to the general population. There can be small population groups in the general population who have high prevalence of deaths by none and CVD deaths.

In the general population there are individuals with genetic lipid disorders. Example of these genetic diseases can be LPL mutations that cause primary hypertriglyceridemia (112, 113). Our study population is compared to a general population where it can be population groups that have higher risk for CVD and none CVD mortality, than the general population.

Methodological limitations in the study design

In none placebo controlled studies there is a possibility for confounding (114). Risk factors like diabetes, hypertension and obesity can have influenced our study population and the general population. Risk factors could have influenced these two populations in different ways. This could have given our study misleading results.

About 15000 individuals have FH in Norway (33). In 2010, 4688 individuals were

registered in the FH registry. Our study population is not a random sample of a population. It is possible that selection bias has occurred (115).

There are different ways for screening FH. To genetically screen the entire population in a country is not cost effective. The most cost effective way to detect FH is by cascade screening (22). This means to investigate first degree relatives of an already diagnosed FH individual.

To diagnose FH, a premature ischemic heart disease history must be present in the case or a close relative (32).

The mortality analysis for CVD mortality in the periods 1992-2005 and 2006-2010 suggest that the period 2006-2010 has a higher total SMR, than the period 1992-2005. In the period 1992-1999 the linked registry had 9 observed deaths. Of these observed deaths, four died of CVD.

Mortality for individuals with familial hypercholesterolemia (FH) in the period 1992 to 2010

Mortality for individuals with familial

hypercholesterolemia (FH) in the period 1992 to 2010

Mortality for individuals with familial

hypercholesterolemia (FH) in the period 1992 to 2010

ABSTRACT:

ACKNOWLEDGEMENTS

CONTENTS

List of Abbreviations

1. BACKGROUND

1.2 Familial hypercholesterolemia 1.2.1 Genetics

1.2.2 Frequency

1.2.3 Diagnosis

1.2.4 Management of FH

1.3

Methodological problems in FH mortality

1.3.1 Challenges in assessment of FH mortality

1.4 FH and death causes

1.5 Aims of the study

2. SUBJECTS AND METHODS

3. RESULTS

4. DISCUSSION

4.2 Methodological consideration